Re: AFIS search candidate list retention poll
Posted: Wed Aug 12, 2009 3:33 pm
ER,
I suspect the candidate parameters could still be adjusted to different numbers, however are constrained by management to the numbers you state.
I recall one of my colleagues getting a hit on candidate #29. I used to kid him that if he only coded his latents better, hits would be higher on the list. In reality, placement of a correct candidate in a list is dependent on several variables, including system capability, quality of the database and coding, as well as skill of the examiner/operator.
I have no doubt that search & matching algorithms have improved and been refined, however many databases still contain converted inked fingerprints from the 1970's and prior. Has the coding on those converted inked fingerprints been upgraded along with system upgrades?
I suspect the candidate parameters could still be adjusted to different numbers, however are constrained by management to the numbers you state.
I recall one of my colleagues getting a hit on candidate #29. I used to kid him that if he only coded his latents better, hits would be higher on the list. In reality, placement of a correct candidate in a list is dependent on several variables, including system capability, quality of the database and coding, as well as skill of the examiner/operator.
I have no doubt that search & matching algorithms have improved and been refined, however many databases still contain converted inked fingerprints from the 1970's and prior. Has the coding on those converted inked fingerprints been upgraded along with system upgrades?